The implication, of course, is that they don’t trust us. Why build a car that commandeers the brakes and wheel if not to eliminate that pesky statistical variable known as “human error”—which is to say, the fallibility that makes us all kin? The sooner we accept this truth the better: in Canada alone, 2,100 people die in traffic accidents each year. Fully 90 per cent of accidents are attributed to driver error.
But there’s something about piloting an automobile that stirs the inner irrationalist, which might explain the glee YouTube viewers have taken in video captured during a recent car safety demonstration in Gothenburg, Sweden. The exhibition was supposed to show off the vaunted “city safety” feature on Volvo’s S60 sedan, which applies brakes automatically in the event of an imminent crash.
Instead, a dais full of journalists was treated to the spectacle of the shiny, orange sedan plowing headlong into the back of a strategically placed transport trailer, then bouncing back after impact with its windshield wipers flapping ridiculously. Erik Coelingh, technical leader of Volvo’s so-called “active safety” program, recently told Maclean’s that the braking system had failed due to lack of power from an improperly charged battery. But there was no avoiding the tsunami of ridicule this sort of footage tends to elicit. “What’s the problem here?” snickered one Web commenter. “It came to a complete stop, no?”
The smart alecks should lap up the schadenfreude while they can. With everyone from established carmakers to the world’s leading Internet search engine testing automated vehicles, the reality of cars doing the driving for us is quickly becoming a reality. How much of it they will be doing, and how soon, are the only unknowns. Google’s much publicized self-driving car has provided ample proof over the past year that automobiles no longer need pilots, while numerous other companies—General Motors, BMW, Volkswagen—have developed cars with varying degrees of autonomy and are planning more in the future.
The technology is advancing more quickly than anyone anticipated. Using a combination of video cameras, radar, light sensors and its own Street View mapping technology, Google’s tricked-out Toyota Priuses had, at last count, travelled nearly 250,000 driverless kilometres in California and Nevada, which have opened their highways to the company for testing (a researcher sits behind the wheel, monitoring the car’s actions on a laptop). The cars can now travel more than 1,400 consecutive kilometres without any form of driver intervention, Google co-founder Sergey Brin revealed last week. The only accident they’ve been involved in took place while one of the cars was under human control.
It’s a stunning achievement, presaging a future when drivers might read, eat or even sleep while their vehicles pilot them to their destinations—and still be safer for it. “Our cars can make decisions about every aspect of driving,” Sebastian Thrun, the leader of the project, told a conference last February. “It is the perfect driving mechanism.”
But that accident—a fender-bender that took place in August near Google’s headquarters in Mountain View, Calif.—highlights the host of issues automation raises. What happens when some cars on the road are robots, and some are not? If a self-driving car causes a crash, who’s responsible? The driver? The carmaker? The designer of the car’s navigation system?
Until those questions are resolved, many researchers believe the spectre of freeways packed with driverless automobiles remains decades away. The scenario now emerging, says Mohan Trivedi, head of the Laboratory for Intelligent and Safe Automobiles at University of California, San Diego, is an interregnum during which cars “assist” drivers by taking on ever more operational functions. “The basic premise is that people do want to stay in their cars and that they enjoy driving,” he says. “But they don’t like to be uncomfortable, inconvenienced or unsafe. So is there something you can do to make the vehicle your assistant?”
The possibilities, says Trivedi, are limitless. Some GM vehicles are now equipped with camera systems that detect nearby vehicles and lane markings, then alert drivers about possible collisions with warning tones and console lights if cars wander off their path. Several car companies have developed systems that detect pedestrians and automatically brake to avoid hitting them and, embarrassing PR glitches aside, Volvo’s “city safety” feature represents the market leader in this area. Using a combination of radar mounted behind the car’s front grille, and digital cameras located under the rear-view mirror, the system first alerts drivers when it detects pedestrians. If the driver fails to respond, it takes over braking altogether.
The same system can prevent rear-enders when the car is travelling up to 30 km/h, and has worked well enough that Volvo plans to expand it to prevent collisions with wild animals. Company technicians recently spent an evening at a safari park gathering digital footage of moose and deer in order to acquaint car computers with their behavioural patterns. With the help of infrared sensors, cars will soon be able to identify creatures, warn drivers and, if necessary, brake automatically before hitting the animals.
The technology is already paying off, according to figures released by the U.S. Insurance Institute for Highway Safety. Volvo’s XC60 SUV, which is equipped with “city safety,” was 22 per cent less likely to be the subject of an insurance claim last year than all other mid-size luxury SUVs combined. It was also less likely to be in crashes than other Volvos—suggesting the difference isn’t merely attributable to the company’s famously safety-conscious customers.
Those numbers will go a long way to quell opposition from anti-automation traditionalists, paving the way for more features that allow robot cars to share the road with human-piloted ones. Researchers at the Massachusetts Institute of Technology, for example, have devised algorithms to predict the behaviour of drivers as they near intersections, and embedded them in a system that will take control of a car’s steering, braking and acceleration to prevent an imminent collision. “Fifty years from now, all cars may be autonomous,” says Domitilla Del Vecchio, the MIT engineer who led the project. “But between now and then, we’re looking at a time when some vehicles will be autonomous and some will be human driven. We are trying to understand how that’s going to work.” Del Vecchio’s team has already tested its system on specially adapted Lexus sedans—one step, she says, toward the day when most of the cars on the road will be machine driven.
Are drivers prepared to make that cultural shift? Mimi Sheller, a Drexel University sociologist studying the impact of car automation, thinks the conversion has already begun. “We’re moving away from the ideas of freedom and individuality we used to associate with driving, toward a rhetoric of security and safety and control,” she says from Philadelphia. At the same time, Sheller adds, the advent of mobile communication technology has “converged” in cars, with the effect of making driving look, to many people, like lost time. “The driver could soon be able to sit back, read emails, make phone calls,” she says. “It means a real change in the cultural meaning of automobility. It also means putting a great deal of trust in the machines.”
Trust many motorists will no doubt be glad to extend to their cars—so long as they know the batteries are charged.